Pages

Friday 26 October 2012

Why Are Manchester United Conceding So Many Opening Goals ?

November beckons and Manchester United find themselves in the leading two in the Premiership with 18 points from a possible 24 and top of their UEFA Champions league group with a perfect start from their three matches so far. Neither set of results should be surprising, United will surely challenge for the Premiership title and seeding ensures that they are the major force in their UCL group. However, one aspect of their season so far has attracted comment and SAF was suitably irked by another concession of an opening goal to pause during his denouncement of a "poor" referring decision to refer to his side's tardy starts. In 11 UCL and EPL games so far this year, United have conceded the opening goal in eight matches and The Boss wasn't pleased.

The rate at which a team scores or concedes the first goal in a contest, should there be one, is strongly related to the proportion of goals they would be expected to score in such match ups. Unsurprisingly United have been pre game favourites in every game they have played so far this year and as such would have also been favourites to claim any opening goal. Their favouritism in the first goal scoring stakes has ranged from slim ( a shade over 50%) verses Liverpool to very strong (over 85%) in their home encounter with Wigan.

Overall United's pre game supremacy suggests a total of eight opening goals for The Red Devils as the most likely outcome from their 11 "league" matches played this season. So by scoring just three they have perfectly emulated the role expected of their inferior opponents. Surely this deviation from expectations is so large as to indicate that something is badly amiss at the heart of their defence?

A team's recorded match result will always be a combination of their overall performance. It's easy to blame the defence for frequently conceding the opening strike, but we also need to include the failure of the attacking side of the ball to score before their opponents. United's forwards were given almost an hour against Everton on opening day to strike before Fellaini's opener, half a game before Gerrard's opener at Anfield and in a similar vein the attack's task was made much easier when the defence kept Wigan, Galatasary and Newcastle goalless for the entire game. Whether a team gives up the first goal is really a joint team effort.

The same is also true when it comes to retrieving losing positions. It's obviously essential that the team scores in response to falling behind, but the value of those goals are also dependent upon the defence remaining firm. Below I've listed the likely range of outcomes faced by Manchester United in the eight matches where they fell behind.

United's Record When Conceding First In 2012/13.

Games Where United Fell Behind. United's % Chance of Winning from 0-1. United's % Chance of Drawing from 0-1.
v Everton. 12 27
v Fulham. 49 25
v Southampton. 34 27
v Liverpool. 26 32
v Tottenham. 37 26
v Stoke. 47 26
v Cluj. 30 27
v Braga. 47 26
Cumulative Expected Success Rate.
0.49
Actual Success Rate.
0.75

On average a the team such as United would emerge from these losing positions with a success rate that combines wins and draws of around 50%. The reality was that United managed a success rate of 75% made up of 6 wins and no draws from the 8 matches. The strikeforce naturally grabbed the headlines, but the impressive record also owed something to the defence preventing further decisive goals, most notably away at Anfield and against Cluj. So just as the attack must share part of the fault for United's tendency to trail in 2012/13, the defence should also share some of the credit for effectively rectifying such situations.

Manchester United's defensive performance may have been judged on high profile incidents rather than overall competence, but this is only part of the reason why their defence has perhaps been too harshly classed as disappointing so far.

The major culprit is sample size.

Eleven games contains a fair amount of data, but the identity of the scorer of the first goal only uses one data point. Therefore, we have a very small sample size which can lead to extreme results and often these extreme results can be plausibly explained by a convincing, but spurious narrative. "United are missing Vidic", for example (conveniently forgetting that Vidic also missed a large proportion of last season, when United's first goal concession rate was as impressive as you would expect, 7/38 games).

The reality is that we are simply seeing media attention being directed onto a luck driven artifact of our small sample size. Few headlines are made if teams such as United go on short term runs where they outscore their first goal expectations, (excellence and improvement appears hardly surprising when displayed by the very best), but when the reverse happens (through chance) then people take notice and begin to try to rationally dissect the merely random.

Two opening goals conceded by United have come about through own goals and however you chose to define randomness or luck, few would expect Rooney to continue to hone his scoring skills by warming up with an initial strike at the wrong end as he did last Saturday against Stoke. A team's talent defines how impressive their long term performances will be, but short term streaks are partly out of their control.

United's Defence. More Concentration or Just More Of The Same ?

We can't know exactly how Manchester United's scoring and conceding patterns will change over the remainder of the season, but long term they will perform close to their historical average. We can guess that the press will speculate that SAF has addressed his defences "problem" of allowing the first goal. And in all likelihood the "improvement" will be attributed to skilled management rather than the more likely absence of bad luck.

For a snapshot of what may await United, we just have to look at their cross City rivals. When I wrote this last season, Manchester City had conceded the first goal in their last five Premiership matches, starting with Swansea and ending with Arsenal . If we also include their two Europa League matches against Sporting Lisbon, the run stretched to seven consecutive contests. A mighty fall from earlier in the season when the Champions elect had been impeccable.

As nothing substantial had occurred that obviously indicated that City were a markedly worse defensive or indeed attacking team than they had been earlier in the year, I suggested that simple luck had been predominantly responsible for the unsatisfactory outcomes and they would probably soon produce results closer to their earlier, more extended run. Happily for their fans, City then opened the scoring in all of their remaining matches.

City's lumpy sequence of partly random, partly skill based results demonstrate the type of forces that are currently shaping United's scoring pattern. Just as importantly, City's run of seven first goal "failures" followed immediately by six consecutive "successes" appear far from random......but they are. Randomness always contains runs of consecutive, identical outcomes, even if we think it shouldn't.

Roy Keane suggested at halftime on Tuesday that United needed to concentrate more and SAF indicated that their problem of conceding first would be dealt with. No doubt steps will be taken in training, but given the likely cause of the run, (luck), doing nothing will probably bring about the same improvement. Starting at Chelsea on Saturday, when, for the first time this season United won't be the favoured team to score first.

1 comment:

  1. "The reality is that we are simply seeing media attention being directed onto a luck driven artifact of our small sample size. Few headlines are made if teams such as United go on short term runs where they outscore their first goal expectations, (excellence and improvement appears hardly surprising when displayed by the very best), but when the reverse happens (through chance) then people take notice and begin to try to rationally dissect the merely random."

    I don't think I can agree with this. I think your assumption that the start they have had is "through chance" is no more likely than the start being a symptom of something they are actually doing wrong. What were the odds that 8 out of 11 games they give up the first goal? And, as you pointed out, if they scored first in all 11, it wouldn't be nearly as surprising due to your probability values. However, we would likely still marvel at it.

    ReplyDelete